Grok Imagine video

grok-imagine-video-1.5 — text-to-video with optional reference images, billed per output second, on the same asynchronous task API as Seedance.

grok-imagine-video-1.5 is xAI's video model, served over the same asynchronous task API as Seedance but with its own request contract. It is text-to-video with optional reference images: there are no first / last frames, no reference video or audio, and no asset library. It is served by the Grok 视频 plan group; a key bound to that group can call it.

Parameters

ParameterDescription
modelgrok-imagine-video-1.5 (required, exact)
prompt or contentOne is required. content accepts text elements only — an image_url, video_url, or audio_url element is rejected at submission
resolutionRequired, unlike Seedance. 480p, 720p, or 1080p. There is no 4K tier
durationOptional integer 1-15 seconds. -1 (automatic) is not accepted. Omit it and HopBase sends nothing, leaving the upstream default in force
aspect_ratioOptional: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3. Note the field name — Seedance's ratio is rejected here, and 21:9 and adaptive are not supported
reference_imagesOptional top-level array of { "url": "https://..." }. URLs must be publicly reachable http(s); data: base64 and asset:// IDs are not accepted

generate_audio, watermark, return_last_frame, frames, seed, camera_fixed, and draft are rejected at submission rather than silently dropped, so a request never appears to have taken effect when it did not.

curl https://api.hop-base.com/v1/video/generate \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "grok-imagine-video-1.5",
    "prompt": "An orange cat runs across sunlit grass as the camera follows",
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "16:9"
  }'

Poll the returned task ID exactly as you would a Seedance task, with GET /v1/video/tasks/{task_id}. status is pending, processing, completed, or failed; treat any non-terminal value as still running. A completed task carries duration_seconds and HopBase relay URLs in outputs. Grok tasks report no token usage and never return last_frame_url.

Video is served by the Grok 视频 plan group, separate from the Grok 全系 group that serves chat and image. A key bound to one does not reach the other. For the Grok chat, reasoning, and image models — including long-context pricing, reasoning-token billing, and server-side tool charges — see Grok.

Rates

Official list prices are on the pricing page. Your own rate is the one shown in the signed-in model catalog, and each request's actual charge is in your usage record.

On this page